Open to opportunities

Deepanshu Yadav

@deepanshuyadav1

Message

AI Research Intern building production RAG and real-time AI services with FastAPI on GCP.

India

Message

What I'm looking for

I’m looking to work on real-time AI systems—RAG pipelines, LLM apps, and backend services on GCP—where I can ship production deployments, control latency/cost, and iterate on model + data workflows.

I built a production RAG pipeline at Indixpert Technology, deploying Vertex AI embeddings with pgvector on Cloud Run and serving real-time AI responses through 15+ async REST endpoints.

To keep costs and latency down, I added a semantic caching layer to reduce redundant Gemini API requests and replaced Vertex AI Vector Search with pgvector on Cloud SQL.

I also implemented an auto-escalation pipeline where sentiment scoring triggers P1 ticket creation and agent handoff, designed for zero human triage.

Outside that work, I’ve shipped AI apps like Langalytics for conversational CSV analysis and a YOLOv8-based traffic management system for detection, tracking, and counting.

Experience

Work history, roles, and key accomplishments

AI Research Intern

Indixpert Technology

May 2026 - Jul 2026 (2 months)

Designed and deployed a production RAG pipeline using Vertex AI embeddings, pgvector, and Gemini 2.5 Flash on Cloud Run with FastAPI. Implemented semantic caching and auto-escalation pipeline with real-time sentiment scoring.